From: terje.mathisen@nospicedham.tmsw.no   
      
   aen@nospicedham.spamtrap.com wrote:   
   > On Thu, 23 Nov 2017 11:26:53 +0100, Terje Mathisen   
   >  wrote:   
   >> ...   
   >> Can you please post some (pseudo) code that explains the algorithm you   
   >> are using?   
   >>   
   > This is the algorithm from TAOCP Vol. 4A, Page 336.   
   > (Sorry the reference to Vol 2 was wrong).   
      
   Thanks, this algorithm is a lot nicer for a SIMD implementation than the   
   one I have been using, even if it does generate more repeates that has   
   to be skipped.   
      
   > ###   
   > Algorithm C (Permutation generation by cyclic shifts). This algorithm   
   > visits all permutations a_1 ... a_n of the distinct elements {x_1,...,   
   > x_n}.   
   >   
   > C1. [Initialize.] Set a_j <- x_j for 1 <= j <= n.   
   > C2. [Visit.] Visit the permutation a_1 ... a_n, and set k <- n.   
   > C3. [Shift.] Replace a_1a_2...a_k& by the cyclic shift a_2...a_ka_1,   
   > and return to C2 if a_k != x_k.   
   > C4. [Decrease k.] Set k <— k — 1, and go back to C3 if k > 1. |   
   >   
   > For example, the successive permutations of {1, 2, 3,4} generated when   
   > n = 4 are   
   > 1234, 2341, 3412, 4123, (1234),   
   > 2314, 3142, 1423, 4231, (2314),   
   > 3124, 1243, 2431, 4312, (3124), (1234),   
   > 2134, 1342, 3421, 4213, (2134),   
   > 1324, 3241, 2413, 4132, (1324),   
   > 3214, 2143, 1432, 4321, (3214), (2134), (1234),   
   > ###   
   > When I call the program up with edb I set a breakpoint at visit, and   
   > can conviently watch how it progresses in the register window on the   
   > right, by looking at XMM0.   
   >   
   Neat!   
      
   Terje   
      
      
   --   
   -    
   "almost all programming can be viewed as an exercise in caching"   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)